Enum HarnessReadRequestV1 

pub enum HarnessReadRequestV1 {
    ContextGet,
    MonitorGet {
        run_id: Option<HarnessRunId>,
    },
    TimelineRead {
        run_id: Option<HarnessRunId>,
        after_sequence: Option<u64>,
        limit: u16,
    },
    TasksList {
        after_task_id: Option<HarnessTaskId>,
        state: Option<HarnessTaskStateV1>,
        parent_task_id: Option<HarnessTaskId>,
        limit: u16,
    },
    TaskGet {
        task_id: HarnessTaskId,
    },
    RunsList {
        task_id: Option<HarnessTaskId>,
        after_run_id: Option<HarnessRunId>,
        lifecycle: Option<HarnessRunLifecycleV1>,
        parent_run_id: Option<HarnessRunId>,
        limit: u16,
    },
    RunGet {
        run_id: HarnessRunId,
    },
    OperationGet {
        operation_id: HarnessOperationId,
    },
    TaskCreate {
        title: String,
        body: String,
        parent_task_id: Option<HarnessTaskId>,
    },
    TaskMove {
        task_id: HarnessTaskId,
        expected_revision: HarnessRevision,
        to: HarnessTaskStateV1,
    },
    RunFinish {
        outcome: HarnessRunFinishOutcomeV1,
        summary: Option<String>,
    },
}

Variants

ContextGet

MonitorGet

Fields

run_id: Option<HarnessRunId>

TimelineRead

Fields

run_id: Option<HarnessRunId>

after_sequence: Option<u64>

limit: u16

TasksList

Fields

after_task_id: Option<HarnessTaskId>

state: Option<HarnessTaskStateV1>

parent_task_id: Option<HarnessTaskId>

“What did this task spawn”: restricts the page to tasks whose own parent_task_id equals this one – direct children only, never deeper descendants. Additive over the original bare { after_task_id, state, limit } shape, so #[serde(default)] keeps an older caller decoding. Composes with state (both filters apply together) and pages the same way via after_task_id.

limit: u16

TaskGet

Fields

task_id: HarnessTaskId

RunsList

Fields

task_id: Option<HarnessTaskId>

after_run_id: Option<HarnessRunId>

lifecycle: Option<HarnessRunLifecycleV1>

parent_run_id: Option<HarnessRunId>

“What did this run spawn”: restricts the page to runs whose own parent_run_id equals this one – direct children only. Additive over the original shape (#[serde(default)]), composes with task_id/lifecycle, pages the same way via after_run_id.

limit: u16

RunGet

Fields

run_id: HarnessRunId

OperationGet

Fields

operation_id: HarnessOperationId

TaskCreate

D5, Slice D: an agent creates a task under its own subtree – parent_task_id: None means “under my own task”; Some must name the caller’s own task or a strict descendant of it (checked service-side, HarnessEngine::task_is_strict_descendant), never a task system-wide the way the operator’s own HarnessCreateTaskRequestV1 can. Gated by grant.task_permissions.create. Every refusal – unreachable parent, a terminal parent, or an invalid title/body/ dependencies – travels back as HarnessTaskCreateResultV1’s own named variant, never a host error. This bound is the coarse wire-level ceiling only; the authoritative title/body rule is the engine’s own HarnessTaskV1::validate(), reached when the mutation applies.

Fields

title: String

body: String

parent_task_id: Option<HarnessTaskId>

TaskMove

D5, Slice D: restricted service-side to a STRICT descendant of the caller’s own task – never the task that governs the caller’s own run (HarnessTaskMoveResultV1::TaskIsOwn). Gated by grant.task_permissions.mutate. to names the operator’s own task graph (validate_operator_move in hatchery-harness-service); an illegal transition or a stale expected_revision is a named refusal, never a host error.

Fields

task_id: HarnessTaskId

expected_revision: HarnessRevision

to: HarnessTaskStateV1

RunFinish

S10: a session reports its OWN run finished – resolved from the caller’s own grant binding (actor_run_id), the same field every other agent verb resolves its identity from; there is no run-id argument, so a session can never finish another session’s run. done applies HarnessLifecycleProjectionV1::CompletedReview (hatchery-harness-service), landing the task in Review – never Done, the review gate a session cannot skip. failed applies HarnessLifecycleProjectionV1::Failed and records a retryable HarnessFailureV1. Never gated by any grant permission (unconditional in allowed_tool_ids, the same way g4a_context_get is): every session needs the ability to report its own work finished regardless of what else its grant allows, or nothing ever closes – the exact structural gap this verb exists to close. summary is accepted on the wire for backward compatibility but has no effect: it used to be posted as mail to the task’s own forum, removed 2026-09-17 when the mailbox moved to its own service (see hatchery-harness-service::read::execute_exact_binding_read’s RunFinish handler). Not idempotent: a second call against an already-terminal run is refused by name as HarnessRunFinishResultV1::AlreadyFinished, never silently accepted.

Fields

outcome: HarnessRunFinishOutcomeV1

summary: Option<String>

Implementations

impl HarnessReadRequestV1

pub fn validate(&self) -> Result<(), HarnessReadApiError>

Trait Implementations

impl Clone for HarnessReadRequestV1

fn clone(&self) -> Self

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for HarnessReadRequestV1

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for HarnessReadRequestV1

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Eq for HarnessReadRequestV1

impl PartialEq for HarnessReadRequestV1

fn eq(&self, other: &Self) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl Serialize for HarnessReadRequestV1

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl StructuralPartialEq for HarnessReadRequestV1